Classification of Packaged Vegetable Soybeans Based on Freshness by Metabolomics Combined with Convolutional Neural...

Area of Science:

• Agricultural Science
• Food Science
• Biotechnology

Background:

• Maintaining vegetable soybean freshness is crucial for quality and reducing post-harvest losses.
• Traditional preservation methods often fall short in extending shelf life effectively.
• Understanding metabolic changes associated with freshness is key to developing advanced preservation strategies.

Purpose of the Study:

• To evaluate the effectiveness of modified atmosphere (MA) packaging in preserving vegetable soybean freshness.
• To investigate the metabolic changes occurring in vegetable soybeans under MA conditions.
• To develop a data-driven model for accurate freshness classification using metabolomics.

Main Methods:

• Vegetable soybeans were stored under modified atmosphere (low O2, high CO2) and normoxic conditions.
• Freshness was assessed by monitoring surface hue angle and analyzing key metabolite concentrations.
• Convolutional neural networks (CNNs) were employed to classify freshness levels based on 62 metabolites.

Main Results:

• MA packaging significantly slowed respiration and reduced the decomposition of pectin and fatty acids.
• Key metabolic pathways, including succinic acid oxidation, were less active under MA conditions, indicating preserved freshness.
• CNNs achieved 92.9% accuracy in classifying vegetable soybean freshness, outperforming linear discriminant analysis by 14.3%.

Conclusions:

• Modified atmosphere packaging effectively extends the freshness of vegetable soybeans by modulating metabolic activity.
• Metabolomic profiling combined with machine learning offers a powerful tool for objective freshness assessment.
• This research provides a foundation for further studies on the metabolism-freshness relationship in horticultural crops.